3rd Symposium on Activity-Based Modeling
After the successful symposia on activity-based models in Seeon, Germany (2022) and Austin, TX (2023), we host another symposium on the advancement of activity-based models in Germany. The symposium will take place 11-13 December 2024 at the TUM Science & Study Center in Raitenhaslach. The host is the research group Travel Behavior at TUM, led by Rolf Moeckel.
Theme of Symposium
The focus of this symposium will be the Representation of Evolutionary Travel Behavior. Of particular interest is the change of travel behavior, triggered by life events, travel time changes, new information, the built environment, policies, pandemics, etc.
Background
Travel behavior is driven – to a large degree – by habitual choices (Kitamura & van der Hoorn 1987). Travel behavior may vary considerably from day to day (Raux et al. 2016), but it does not change dramatically from year to year (McCarthy 1982). Life events, such as household relocation, graduation from school, change of job, birth of a child, a global pandemic, and so forth, may change travel behavior fundamentally. But for most people, such changes are not routine.
Most travel demand models recreate travel behavior from scratch with every model run and ignore habitual behavior. Consequently, travel demand models generally overestimate the impact of scenarios, at least in the first few years. Agent-based models, however, allow remembering the travel choice of the previous simulation period. Thereby, agent-based models enable analysts to explicitly account for habitual behavior. Activity-based models, a subset of agent-based models, generate temporally and spatially explicit activity schedules. Such individual schedules may be adjusted incrementally (or remain unchanged) between two simulation years.
Call for action
Initial steps towards an evolutionary perspective of travel behavior change include cross-sectional data that have been used for longitudinal analysis (Habib & Weis 2014). Evolutionary car ownership models were built (Jong et al. 2014) and several sociodemographic and land use models simulate changes over time (Eluru et al., 2008; Wegener 2014). The impact of habit on travel behavior was quantified (Goulet-Langlois et al. 2018) and activity-based models started to recognize simulated choices from previous simulated years (Auld & Mohammadian 2012, Moeckel et al. 2023). It is time to bring together behavioral researchers, data scientists and modelers to make broad steps towards evolutionary activity-based models.
Topics
Topics that we like to develop at this symposium include:
- How stable is travel behavior over time? How stable are attitudes?
- Now that we have week-long surveys, does that open opportunities to identify habits, and how stable are habits?
- What can we learn from longitudinal data sources and estimation?
- What has changed about travel behavior post-pandemic, and what will stick?
- What would an evolutionary model instead of an equilibrium model look like? This is common for land-use modeling already, how does it translate to transport modeling?
The symposium will improve our understanding of theoretical approaches, data analysis and model design. Contributions may be conceptual (a smart idea that is not proven to work yet), technical (innovative model estimations or simulation solutions) or practical (new implementation to solve real-world problems).
How to Attend
To keep the size of the Symposium on the smaller side, participation is by invitation only. Participants are required to fund their own travel to Munich. Travel between Munich and the venue of the Symposium in Raitenhaslach, all meals and hotel nights from Sunday through Thursday are covered by the host.
To be considered for participation, please submit an abstract describing your presentation.